Releases: sedthh/lara-hungarian-nlp
Releases · sedthh/lara-hungarian-nlp
v2.0.0
v1.2.0
- parser Intents() match_order() now accepts preferences in reverse order
- typo_prefix and typo_affix now accepts matches where the prefixes/affixes are separated with a space from the stem
- fixed a bug where an empty typo_stem would match all texts
- parser Extract() commands() now returns a single list and allows quotes to define keywords with spaces in them
- parser Extract() number converter now accepts numbers like "háromszor" "kétszer"
- fixed a bug where \b would be interpreted as a character instead of \b for regex boundary declarations in stems
- updated inverse stemmer for articles and added the word "Budapest" as an exception
- nlp now has a function for generating "a" or "az"
- parser Intents now matches typo stems with accidental "ű" letters pressed at the end
- fixed a bug where parser Extract() timestamps() multiple times would return an empty list (due to caching of regex finditer functions)
- fixed reported issue with dates()
- fixed reported issue with times()
- updated entities based on user feedback
- updated and added further topics to smalltalk entities
- added more test cases and examples
- updated README.md
1.1.6
- added just_saking() stemmer for stemming words in questions
- added inverse() stemmer for adding prefixes to words in accordance with their vowel harmony
- improved vowel harmony detection
- added match_order() function for Parser Class to match Intents based on preferences
- added more test cases
- updated some examples